The present invention relates to cables made of twisted conductor pairs. More specifically, the present invention relates to twisted pair cables for high-speed data communications applications.
With the widespread and growing use of computers in communications applications, the ensuing volumes of data traffic have accentuated the need for communications networks to transmit the data at higher speeds. Moreover, advancements in technology have contributed to the design and deployment of high-speed communications devices that are capable of communicating the data at speeds greater than the speeds at which conventional data cables can propagate the data. Consequently, the data cables of typical communications networks, such as local area network (LAN) communities, limit the speed of data flow between communications devices.
In order to propagate data between the communications devices, many communications networks utilize conventional cables that include twisted conductor pairs (also referred to as “twisted pairs” or “pairs”). A typical twisted pair includes two insulated conductors twisted together along a longitudinal axis.
The twisted pair cables must meet specific standards of performance in order to efficiently and accurately transmit the data between the communication devices. If cables do not at least satisfy these standards, the integrity of their signals is jeopardized. Industry standards govern the physical dimensions, the performance, and the safety of the cables. For example, in the United States, the Electronic Industries Association/Telecommunications Industry Association (EIA/TIA) provides standards regarding the performance specifications of data cables. Several foreign countries have also adopted these or similar standards.
According to the adopted standards, the performance of twisted pair cables is evaluated using several parameters, including dimensional properties, interoperability, impedance, attenuation, and crosstalk. The standards require that the cables perform within certain parameter boundaries. For instance, a maximum, average outer cable diameter of 0.250″ is specified for many twisted pair cable types. The standards also require that the cables perform within certain electrical boundaries. The range of the parameter boundaries varies depending on the attributes of the signal to be propagated over the cable. In general, as the speed of a data signal increases, the signal becomes more sensitive to undesirable influences from the cable, such as the effects of impedance, attenuation, and crosstalk. Therefore, high-speed signals require better cable performance in order to maintain adequate signal integrity.
A discussion of impedance, attenuation, and crosstalk will help illustrate the limitations of conventional cables. The first listed parameter, impedance, is a unit of measure, expressed in Ohms, of the total opposition offered to the flow of an electrical signal. Resistance, capacitance, and inductance each contribute to the impedance of a cable's twisted pairs. Theoretically, the impedance of the twisted pair is directly proportional to the inductance from conductor effects and inversely proportional to the capacitance from insulator effects.
Impedance is also defined as the best “path” for data to traverse. For instance, if a signal is being transmitted at an impedance of 100 Ohms, it is important that the cabling over which it propagates also possess an impedance of 100 Ohms. Any deviation from this impedance match at any point along the cable will result in reflection of part of the transmitted signal back towards the transmission end of the cable, thereby degrading the transmitted signal. This degradation due to signal reflection is known as return loss.
Impedance deviations occur for many reasons. For example, the impedance of the twisted pair is influenced by the physical and electrical attributes of the twisted pair, including: the dielectric properties of the materials proximate to each conductor; the diameter of the conductor; the diameter of the insulation material around the conductor; the distance between the conductors; the relationships between the twisted pairs; the twisted pair lay lengths (distance to complete one twist cycle); the overall cable lay length; and the tightness of the jacket surrounding the twisted pairs.
Because the above-listed attributes of the twisted pair can easily vary over its length, the impedance of the twisted pair may deviate over the length of the pair. At any point where there is a change in the physical attributes of the twisted pair, a deviation in impedance occurs. For example, an impedance deviation will result from a simple increase in the distance between the conductors of the twisted pair. At the point of increased distance between the twisted pairs, the impedance will increase because impedance is known to be directly proportional to the distance between the conductors of the twisted pair.
Greater variations in impedance will result in worse signal degradation. Therefore, the allowable impedance variation over the length of a cable is typically standardized. In particular, the EIA/TIA standards for cable performance require that the impedance of a cable vary only within a limited range of values. Typically, these ranges have allowed for substantial variations in impedance because the integrity of traditional data signals has been maintained over these ranges. However, the same ranges of impedance variations jeopardize the integrity of high-speed signals because the undesirable effects of the impedance variations are accentuated when higher speed signals are transmitted. Therefore, accurate and efficient transmissions of high-speed signals, such as signals with aggregate speeds approaching and surpassing 10 gigabits per second, benefit from stricter control of the impedance variations over the length of a cable. In particular, post-manufacture manipulations of a cable, such as twisting the cable, should not introduce significant impedance mismatches into the cable.
The second listed parameter useful for evaluating cable performance is attenuation. Attenuation represents signal loss as an electrical signal propagates along a conductor length. A signal, if attenuated too much, becomes unrecognizable to a receiving device. To make sure this doesn't happen, standards committees have established limits on the amount of loss that is acceptable.
The attenuation of a signal depends on several factors, including: the dielectric constants of the materials surrounding the conductor; the impedance of the conductor; the frequency of the signal; the length of the conductor; and the diameter of the conductor. In order to help ensure acceptable attenuation levels, the adopted standards regulate some of these factors. For example, the EIA/TIA standards govern the allowable sizes of conductors for the twisted pairs.
The materials surrounding the conductors affect signal attenuation because materials with better dielectric properties (e.g., lower dielectric constants) tend to minimize signal loss. Accordingly, many, conventional cables use materials such as polyethylene and fluorinated ethylene propylene (FEP) to insulate the conductors. These materials usually provide lower dielectric loss than other materials with higher dielectric constants, such as polyvinyl chloride (PVC). Further, some conventional cables have sought to reduce signal loss by maximizing the amount of air surrounding the twisted pairs. Because of its low dielectric constant (1.0), air is a good insulator against signal attenuation.
The material of the jacket also affects attenuation, especially when a cable does not contain internal shielding. Typical jacket materials used with conventional cables tend to have higher dielectric constants, which can contribute to greater signal loss. Consequently, many conventional cables use a “loose-tube” construction that helps distance the jacket from unshielded twisted pairs.
The third listed parameter that affects cable performance is crosstalk. Crosstalk represents signal degradation due to capacitive and inductive coupling between the twisted pairs. Each active twisted pair naturally produces electromagnetic fields (collectively “the fields” or “the interference fields”) about its conductors. These fields are also known as electrical noise or interference because the fields can undesirably affect the signals being transmitted along other proximate conductors. The fields typically emanate outwardly from the source conductor over a finite distance. The strengths of the fields dissipate as the distances of the fields from the source conductor increase.
The interference fields produce a number of different types of crosstalk. Near-end crosstalk (NEXT) is a measure of signal coupling between the twisted pairs at positions near the transmitting end of the cable. At the other end of the cable, far-end crosstalk (FEXT) is a measure of signal coupling between the twisted pairs at a position near the receiving end of the cable. Powersum crosstalk represents a measure of signal coupling between all the sources of electrical noise within a cable entity that can potentially affect a signal, including multiple active twisted pairs. Alien crosstalk refers to a measure of signal coupling between the twisted pairs of different cables. In other words, a signal on a particular twisted pair of a first cable can be affected by alien crosstalk from the twisted pairs of a proximate second cable. Alien Power Sum Crosstalk (APSNEXT) represents a measure of signal coupling between all noise sources outside of a cable that can potentially affect a signal.
The physical characteristics of a cable's twisted pairs and their relationships to each other help determine the cable's ability to control the effects of crosstalk. More specifically, there are several factors known to influence crosstalk, including: the distance between the twisted pairs; the lay lengths of the twisted pairs; the types of materials used; the consistency of materials used; and the positioning of twisted pairs with dissimilar lay lengths in relation to each other. In regards to the distance between the twisted pairs of the cable, it is known that the effects of crosstalk within a cable decrease when the distance between twisted pairs is increased. Based on this knowledge, some conventional cables have sought to maximize the distance between each particular cable's twisted pairs.
In regards to the lay lengths of the twisted pairs, it is generally known that twisted pairs with similar lay lengths (i.e., parallel twisted pairs) are more susceptible to crosstalk than are non-parallel twisted pairs. This increased susceptibility to crosstalk exists because the interference fields produced by a first twisted pair are oriented in directions that readily influence other twisted pairs that are parallel to the first twisted pair. Based on this knowledge, many conventional cables have sought to reduce intra-cable crosstalk by utilizing non-parallel twisted pairs or by varying the lay lengths of the individual twisted pairs over their lengths.
It is also generally known that twisted pairs with long lay lengths (loose twist rates) are more prone to the effects of crosstalk than are twisted pairs with short lay lengths. Twisted pairs with shorter lay lengths orient their conductors at angles that are farther from parallel orientation than are the conductors of long lay length twisted pairs. The increased angular distance from a parallel orientation reduces the effects of crosstalk between the twisted pairs. Further, longer lay length twisted pairs cause more nesting to occur between pairs, creating a situation where distance between twisted pairs is reduced. This further degrades the ability of pairs to resist noise migration. Consequently, the long lay length twisted pairs are more susceptible to the effects of crosstalk, including alien crosstalk, than are the short lay length twisted pairs.
Based on this knowledge, some conventional cables have sought to reduce the effects of crosstalk between long lay length twisted pairs by positioning the long lay length pairs farthest apart within the jacket of the cable. For example, in a 4-pair cable, the two twisted pairs with the longer lay lengths would be positioned farthest apart (diagonally) from each other in order to maximize the distance between them.
With the above cable parameters in mind, many conventional cables have been designed to regulate the effects of impedance, attenuation, and crosstalk within individual cables by controlling some of the factors known to influence these performance parameters. Accordingly, conventional cables have attained levels of performance that are adequate only for the transmission of traditional data signals. However, with the deployment of emerging high-speed communications systems and devices, the shortcomings of conventional cables are quickly becoming apparent. The conventional cables are unable to accurately and efficiently propagate the high-speed data signals that can be used by the emerging communications devices. As mentioned above, the high-speed signals are more susceptible to signal degradation due to attenuation, impedance mismatches, and crosstalk, including alien crosstalk. Moreover, the high-speed signals naturally worsen the effects of crosstalk by producing stronger interference fields about the signal conductors.
Due to the strengthened interference fields generated at high data rates, the effects of alien crosstalk have become more significant to the transmission of high-speed data signals. While conventional cables could overlook the effects of alien crosstalk when transmitting traditional data signals, the techniques used to control crosstalk within the conventional cables do not provide adequate levels of isolation to protect from cable to cable alien crosstalk between the conductor pairs of high-speed signals. Moreover, some conventional cables have employed designs that actually work to increase the exposure of their twisted pairs to alien crosstalk. For example, typical star-filler cables often maintain the same cable diameter by reducing the thickness of their jackets and actually pushing their twisted pairs closer to the jacket surface, thereby worsening the effects of alien crosstalk by bringing the twisted pairs of proximate conventional cables closer together.
The effects of powersum crosstalk are also increased at higher data transmission rates. Traditional signals such as 10 megabits per second and 100 megabits per second Ethernet signals typically use only two twisted pairs for propagation over conventional cables. However, higher speed signals require increased bandwidth. Accordingly, high-speed signals, such as 1 gigabit per second and 10 gigabits per second Ethernet signals, are usually transmitted in full-duplex mode (2-way transmission over a twisted pair) over more than two twisted pairs, thereby increasing the number of sources of crosstalk. Consequently, conventional cables are not capable of overcoming the increased effects of powersum crosstalk that are produced by high-speed signals. More importantly, conventional cables cannot overcome the increases of cable to cable crosstalk (alien crosstalk), which crosstalk is increased substantially because all of the twisted pairs of adjacent cables are potentially active.
Similarly, other conventional techniques are ineffective when applied to high speed communications signals. For example, as mentioned above, some traditional data signals typically need only two twisted pairs for effective transmissions. In this situation, communications systems can usually predict the interference that one twisted pair's signal will inflict on the other twisted pair's signal. However, by using more twisted pairs for transmissions, complex high-speed data signals generate more sources of noise, the effects of which are less predictable. As a result, conventional methods used to cancel out the predictable effects of noise are no longer effective. In regards to alien crosstalk, predictability methods are especially ineffective because the signals of other cables are usually unknown or unpredictable. Moreover, trying to predict signals and their coupling effects on adjacent cables is impractical and difficult.
The increased effects of crosstalk due to high-speed signals pose serious problems to the integrity of the signals as they propagate along conventional cables. Specifically, the high-speed signals will be unacceptably attenuated and otherwise degraded by the effects of alien crosstalk because conventional cables traditionally focus on controlling intra-cable crosstalk and are not designed to adequately combat the effects of alien crosstalk produced by high-speed signal transmissions.
Conventional cables have used traditional techniques to reduce intra-cable crosstalk between twisted pairs. However, conventional cables have not applied those techniques to the alien crosstalk between adjacent cables. For one, conventional cables have been able to comply with specifications for slower traditional data signals without having to be concerned with controlling alien crosstalk. Further, suppressing alien crosstalk is more difficult than controlling intra-cable cross-talk because, unlike intra-cable crosstalk from known sources, alien crosstalk cannot be precisely measured or predicted. Alien crosstalk is difficult to measure because it typically comes from unknown sources at unpredictable intervals.
As a result, conventional cabling techniques have not been successfully used to control alien crosstalk. Moreover, many traditional techniques cannot be easily used to control alien crosstalk. For example, digital signal processing has been used to cancel out or compensate for effects of intra-cable crosstalk. However, because alien crosstalk is difficult to measure or predict, known digital signal processing techniques cannot be cost effectively applied. Thus, there exists an inability in conventional cables to control alien crosstalk.
In short, conventional cables cannot effectively and accurately transmit high-speed data signals. Specifically, the conventional cables do not provide adequate levels of protection and isolation from impedance mismatches, attenuation, and crosstalk. For example, the Institute of Electrical and Electronics Engineers (IEEE) estimates that in order to effectively transmit 10 Gigabit signals at 100 megahertz (MHz), a cable must provide at least 60 dB of isolation against noise sources outside of the cable, such as adjacent cables. However, conventional cables of twisted conductor pairs typically provide isolations well short of the 60 dB needed at a signal frequency of 100 MHz, usually around 32 dB. The cables radiate about nine times more noise than is specified for 10 Gigabit transmissions over a 100 meter cabling media. Consequently, conventional twisted pair cables cannot transmit the high-speed communications signals accurately or efficiently.
Although other types of cables have achieved over 60 dB of isolation at 100 MHz, these types of cables have shortcomings that make their use undesirable in many communications systems, such as LAN communities. A shielded twisted pair cable or a fiber optic cable may achieve adequate levels of isolation for high-speed signals, but these types of cables cost considerably more than unshielded twisted pairs. Unshielded systems typically enjoy significant cost savings, which savings increase the desirability of unshielded systems as a transmitting medium. Moreover, conventional unshielded twisted pair cables are already well-established in a substantial number of existing communications systems. It is desirable for unshielded twisted pair cables to communicate high-speed communication signals efficiently and accurately. Specifically, it is desirable for unshielded twisted pair cables to achieve performance parameters adequate for maintaining the integrity of high-speed data signals during efficient transmission over the cables.